He called the dispute a distraction from superintelligence risks.
I'm guessing: - Anthropic "agrees" to do something minor/performative bc the issue was overstated to begin with - this sort of allows the administration to "save face" (but not really) - Fable gets unblocked - and we just wasted another several days with ASI looming
Many users dismissed Anthropic's expected minor concessions as empty posturing that harms the company more than admitted and changes nothing, while others hoped for better future policy relationships and praised Fable's performance.
When I say "the issue was overstated" -- as I've said elsewhere, I think a plausible argument could be made that Fable/Mythos deployment is bad, but it would also likely apply to GPT-5.5.
When I say "the issue" I mean "the claim that there was some new insight revealed Friday"
@JeffLadish Talking about a manufactured crisis instead of otherwise preparing for inevitable crises, not using Fable for beneficial + research + security purposes, polarizing against one another for no reason
@Miles_Brundage In what way did we waste the days?
@_Jason_Dean_ They already do actual pro bono stuff, like Glasswing, IIUC...
@Miles_Brundage I’m just actually not sure what the admin believes here! There are smart cyber people in OSTP and in the NSA and CAISI. I don’t know who is talking to who and how they’re assessing risk. Transparency here would be super helpful
@Miles_Brundage In what way did we waste the days?
I'm guessing: - Anthropic "agrees" to do something minor/performative bc the issue was overstated to begin with - this sort of allows the administration to "save face" (but not really) - Fable gets unblocked - and we just wasted another several days with ASI looming
@Miles_Brundage Maybe Anthropic does some pro bono (or at-cost) cybersecurity work for the government
@JeffLadish I hope something like that happens but the last crisis did not seem to produce that outcome sadly
@Miles_Brundage Maybe this is naive but I’m hoping they work through this and come out with better working relationships, so next time when there’s more danger they can work together more effectively.
Like the labs and government are going to have to work together! The alternatives are bad
@Miles_Brundage True but I’m hoping this one had a better shot because it’s easier to evaluate claims about cyber risk than it is about contract language for military work
@Miles_Brundage ASI delayed by 2 days and you're blackpilling?
I'm guessing: - Anthropic "agrees" to do something minor/performative bc the issue was overstated to begin with - this sort of allows the administration to "save face" (but not really) - Fable gets unblocked - and we just wasted another several days with ASI looming
@Miles_Brundage the issue is more the short-term implications of model-specific regulatory bans & how the market might interpret & value these during IPO (impacts openAI too). the precedent has been set, so the risk is magnified until 2028 at least
@_Jason_Dean_ But yes they could re-announce it somehow :)
@Miles_Brundage I stayed up for days, burned 2 x20 accounts and got a 3rd. Gave the hardest problems I have that Opus had already maxed out on.
Fable absolutely CRUSHED it, i also had it running scans and building plans. 1000s of tickets because i knew it wasn't going to last
@Miles_Brundage seems right.
i hope the messiness in these pre-ASI days help congress (and others) realize just how urgent it is that we get in place much better policy and staff up accordingly
voices like yours are helping move us in that direction.
@Miles_Brundage this probably punts any legitimate government action further down the line too
@Miles_Brundage translation: nothing changes but everyone gets to claim they won
@Miles_Brundage And the Administration demands the equity.
@Miles_Brundage I can't believe you are paid to shill this "ASI" nonsense.
@Miles_Brundage that timeline math is the part that actually stings
@Miles_Brundage